NBA Attendance Analysis

One of the topics that I talk about in Sports Economics is the impact that a lockout has on attendance.  Obviously, with a lockout the total number of people that attend all league games for that season will fall as there are fewer number of total games played.  Yet that does not really get at the issue at hand.  So, one way to tackle this problem is to use intervention analysis.  In essence, intervention analysis is a time series statistical technique that examines how the average value of a series of numbers changes given a specific intervention.  Specifically, we looked at how the first difference in the average value of total league attendance changed due to a strike or a lockout, given that attendance is non-stationary over the time period that we were looking at. This is exactly what we did in chapter 2 of the Wages of Wins.

Unfortunately, at this point in time, I cannot first difference the attendance data for the current NBA season, since the next NBA season (2012-2013) has not been played yet.  Thus intervention analysis is not currently possible to analyze the impact of the NBA lockout.

So, using total league attendance does not help, since the number of games played is different, and it is too soon to use intervention analysis.  Is there any alternative way to figure this out?  Yes.  One is to perform a t-test on the means of attendance.  So, I downloaded the team-by-team attendance data from ESPN, and entered and sorted the data into Microsoft Excel, and then calculated a t-test on home team average attendance for the 2011-2012 and 2010-2011 NBA teams.  The result of the t-test was that team-by-team average home attendance was not significantly different over the two seasons.  In other words, statistically speaking there was not a great enough difference in the two data series to say that one is different from the other; and thus I would conclude that the NBA lockout did not have a statistically significant impact on average NBA team home attendance between those two seasons.

Alternatively, one could just plot the data over the last few seasons and look to see if there is any visual difference.  While this is not all that technical, it does give some additional evidence that fans did not respond much because of the NBA lockout.  Here is a line graph of average NBA home attendance over the last four NBA seasons.  The line in blue is the 2011-2012 lockout season.


Average NBA Team Home Attendance:  2008/09 to 2011/12

Scoring in the NBA can be Deceptive

Dave Berri has a nice article on how total points scored by NBA players can be deceptive.

USA Today's Latest NCAA Finance Database

The USA Today has published it latest NCAA finance database.

Here is some highlights from the University of Iowa:

	Ticket		Student		Total		Scholar-		Coaching		Building/		Total
Year	Sales		Fees		Revenue		ships		Staff		Grounds		Expenses
2011	$23,180,905		$564,680		$93,353,561		$9,362,572		$29,016,057		$21,863,477		$88,057,486
2010	$21,815,895		$525,707		$88,735,093		$8,585,730		$26,197,937		$14,944,085		$74,438,196
2009	$21,922,358		$525,941		$79,971,143		$8,755,400		$24,453,669		$13,703,314		$71,116,911
2008	$19,103,235		$1,487,795		$81,515,865		$7,579,781		$23,166,018		$16,003,409		$71,602,594
2007	$21,731,819		$1,494,706		$80,832,070		$6,722,602		$21,376,108		$15,161,903		$70,469,589
2006	$20,086,445		$1,495,060		$73,321,227		$6,652,458		$21,063,488		$6,420,286		$59,224,861

NHL Pay and Performance

Last month I looked a pay and performance in the NHL for just the 2011-2012 season. As I mention in the blog, that is a rather small sample. So to correct for this sample size problem, let's take a look over a longer period of time; from the 2000-2001 regular season to the 2011-2012 regular season (no 2004-2005 as this was the cancelled NHL season).  I am following the same payroll and performance analysis that I blogged about with regard to Major League Baseball for this NHL study.

For the eleven seasons covered, relative payroll is positive and statistically significant, which is expected.  On the other hand, relative payroll only "explains" about 25% of team performance as measured by points percent (which is the number of points divided by the maximum number of points possible and is the same as winning percent).  Again, while the relationship between the two variables are what is to be expected, the amount of "bang for the buck" is not so great.

Philly fans

Say what you want about sports fans in Philadelphia - being born and raised in Southeastern Pennsylvania - I have great memories of sports fans in Philly, here is a remarkable reaction of the fans in response to the news announcement that Osama bin Ladin (who declared war on the US) was dead. I was at home watching this game on ESPN, and remember this vividly last year. So on the one year anniversary - here it is.

NHL Goalie Performance Measure

Here is a step-by-step guide to David Berri's and my NHL goalie measure from our article in the Journal of Sports Economics in 2010.

Step 1: Calculate the Marginal Value of a Goal Against. I have previously posted on how to do this, so follow the steps in this blog.
Step 2: Download NHL Goalie Data I downloaded the 2011-2012 NHL goalie data from NHL.com. Go to the web page, click on Stats, then choose Individuals. Once that page opens, I choose 2011-2012 Regular Season and under Position I choose Goalie. This opens up a page with all the NHL goalies for the time period you are analyzing. Make sure you get all the goalie data by choosing the additional tables at the bottom of the first table. I then copied and pasted each table into Microsoft Excel.
Step 3: Wins Above Average (WAA) Calculation. WAA is measured as the absolute value (since a goal against has a negative effect on team wins) of the marginal value of a goal against divided by two (since each win is worth two standings points) times the number of shots on goal that goalie faces times the difference in the save percentage of the goalie and the average save percentage of all goalies for that season.  Let's break this calculation up into three parts.
Part A:  This is already done, since the data is just the columns listed as SA and SV% from NHL.com for each goalie, which in my spreadsheet is column I for SA and column M for SV%. 
Part B:  Calculate the average save percentage for the league by finding the total number of shots against and the total number of goals against and then subtract the total number of goals against from the total number of shots against and divide that by the total number of shots against.  In Excel this would be (S6-S9)/S6, if total shots against is in cell S6 and total goals against is in cell S9.  Suppose this formula is put in cell S12.
Part C:  The first part of the equation [the marginal value of a goal against] was found in step 1.  Take the absolute value of this number and divide by two.  If the marginal value of a goal against is in cell S3, then in Excel type ABS($S$3)/2.
Put it all together and in Excel we have the following formula:  =(M2-$S$12)*I2*(ABS($S$3)/2).  Note that M2 is the first goalie listed save percentage and cell I2 is the first goalie listed shots against.  Then copy and paste this formula down in the spreadsheet for each goalie and you have an estimate of each goalie's Wins Above Average (WAA) for that season.
Step 4:  Rank Most Productive Goalies.  I then rank each goalie by their WAA from highest to lowest for that season.  You can do this by sorting from highest to lowest.  I do something slightly different but it is the same thing, it just allows me to do this without actually having to use the sort option in Excel.

NCAA Basketball Coaches Salary

New York Times piece on the differences in salary of male and female NCAA basketball coaches.